ASIAN DEVELOPMENT BANK
OUTLOOK 2016
ASIA’S POTENTIAL GROWTH
ASIAN DEVELOPMENT BANK
OUTLOOK 2016
www.adb.org; openaccess.adb.org Some rights reserved. Published in 2016. Printed in the Philippines.
ISBN 978-92-9257-385-0 (Print), 978-92-9257-386-7 (e-ISBN) ISSN 0117-0481
Publication Stock No. FLS157678-3
Cataloging-In-Publication Data Asian Development Bank.
Asian development outlook 2016. Asia’s potential growth. Mandaluyong City, Philippines: Asian Development Bank, 2016.
1. Economics. 2. Finance. 3. Asia. I. Asian Development Bank.
The views expressed in this publication are those of the authors and do not necessarily reflect the views and policies of the Asian Development Bank (ADB) or its Board of Governors or the governments they represent.
ADB does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The mention of specific companies or products of manufacturers does not imply that they are endorsed or recommended by ADB in preference to others of a similar nature that are not mentioned.
By making any designation of or reference to a particular territory or geographic area, or by using the term “country” in this document, ADB does not intend to make any judgments as to the legal or other status of any territory or area. This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO)
https://creativecommons.org/licenses/by/3.0/igo/. By using the content of this publication, you agree to be bound by the terms of said license as well as the Terms of Use of the ADB Open Access Repository at openaccess.adb.org/termsofuse
This CC license does not apply to non-ADB copyright materials in this publication. If the material is attributed to another source, please contact the copyright owner or publisher of that source for permission to reproduce it. ADB cannot be held liable for any claims that arise as a result of your use of the material.
Attribution—In acknowledging ADB as the source, please be sure to include all of the following information: Author. Year of publication. Title of the material. © Asian Development Bank [and/or Publisher]. https://openaccess.adb.org. Available under a CC BY 3.0 IGO license.
Translations—Any translations you create should carry the following disclaimer:
Originally published by the Asian Development Bank in English under the title [title] © [Year of publication] Asian Development Bank. All rights reserved. The quality of this translation and its coherence with the original text is the sole responsibility of the [translator]. The English original of this work is the only official version.
Adaptations—Any adaptations you create should carry the following disclaimer:
This is an adaptation of an original Work © Asian Development Bank [Year]. The views expressed here are those of the authors and do not necessarily reflect the views and policies of ADB or its Board of Governors or the governments they represent. ADB does not endorse this work or guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use.
Please contact OARsupport@adb.org or publications@adb.org if you have questions or comments with respect to content, or if you wish to obtain copyright permission for your intended use that does not fall within these terms, or for permission to use the ADB logo.
Foreword
Acknowledgments Definitions Abbreviations
ADO 2016—Highlights
Part 1 Rescuing growth in uncertain times
Slow going in a tough global environment Spillover from the People’s Republic of China Emergent producer price deflation
Annex: A challenging and uncertain global outlook
Part 2 Asia’s potential growth
Understanding the growth slowdown Determinants of potential growth A “new normal” for potential growth? Policies to invigorate potential growth
Part 3 Economic trends and prospects
in developing Asia
Central Asia
Armenia Azerbaijan Georgia Kazakhstan Kyrgyz Republic Tajikistan Turkmenistan UzbekistanEast Asia
People’s Republic of China Hong Kong, China
Republic of Korea Mongolia
India Maldives Nepal Pakistan Sri Lanka
Southeast Asia
Brunei Darussalam Cambodia IndonesiaLao People’s Democratic Republic Malaysia Myanmar Philippines Singapore Thailand Viet Nam
The Pacific
FijiPapua New Guinea Solomon Islands Timor-Leste Vanuatu
North Pacific economies South Pacific economies Small island economies
2
ASIA’S POTENTIAL
GROWTH
This chapter was written by Jesus Felipe, of the Economic Research and Regional Cooperation Department, ADB, Manila, Matteo Lanzafame, of the University of Messina, and Miguel León-Ledesma, of the University of Kent. Contributions from Bart Verspagen and Neil Foster-McGregor of the United Nations University–Maastricht Economic and Social Research Institute on Innovation and Technology, as well as from Noli Sotocinal, of the Economic Research and Regional Cooperation Department are acknowledged. Likewise, comments from participants at a workshop in Seoul on 19–20 November 2015, are also noted.
Many Asian economies have registered remarkably high output growth rates in the past 3 decades. This achievement came about despite disruptions caused by economic crises. Arguably, the most serious downturn was the Asian financial crisis of 1997–1998, which severely affected Southeast Asia and the Republic of Korea. By comparison, the global financial crisis (GFC) of 2008–2009 has had much less impact on Asia. The main reason why the region’s economies were not directly affected is because their exposure to United States (US) subprime mortgages was low. The region was nevertheless affected indirectly as demand for its exports from its main trading partners contracted significantly.
This was enough to disrupt growth momentum in Asia, and as a result economic growth in Asia slowed in the aftermath of the GFC. From an average of 8.3% during 2006–2010, gross domestic product (GDP) growth in the region fell to 5.9% in 2015 (Figure 2.1.1). This downward trend in regional growth follows 6 years of accelerating growth in the interval 2001–2007. The People’s Republic of China (PRC), which accounts for a large share of the region’s GDP, has been the main driver of regional economic growth. Given the recent slowdown in the PRC, some observers have started talking about a “new normal” of substantially lower growth in the region.
While developing Asia is not alone in this regard—growth also declined in Africa, Latin America, and Eastern Europe— its slowdown is likely to have repercussions for the region and the rest of the world. Developing Asia’s success in lifting 1 billion individuals out of poverty during 1990–2012 hinged
on its ability to sustain high rates of economic growth. Moreover, as the region currently accounts for more than a quarter of world GDP as valued at market exchange rates, a persistent slowdown in developing Asia threatens to undermine the fragile global recovery.
This theme chapter aims to reveal the factors underlying the growth slowdown and the extent to which the slowdown reflects changes in the region’s productive capacity. What has happened to developing Asia’s productive capacity—its so-called potential growth—and how can policy makers respond?
2.1.1 Average GDP growth in selected regions
8.2 7.5 6.1 1.6 6.6 7.1 4.9 6.57.1 7.98.4 9.310.1 6.8 6.1 9.3 7.4 6.3 6.4 6.35.9 % −4 0 4 8 12 1995 2000 2005 2010 2015 Africa Eastern Europe Latin America Developing Asia
Understanding the
growth slowdown
The issues posed by the decline in Asia’s growth rate, and the nature of that decline, are important from a policy perspective. The right policy response depends on the nature of the region’s slowdown. Is it a temporary—albeit prolonged—effect of the business cycle? Or are more persistent changes under way? If weaker growth reflects slack demand such as softening export orders or a downturn in private investment, then fiscal or monetary stimulus may be needed for temporary support. But if supply-side factors are at play, and growth moderation stems from slowing expansion of the region’s productive capacity, then any revival of growth prospects may depend on structural policy reform.
The concepts of potential GDP and potential growth, which are grounded in a view of an economy’s productive capacity under conditions of stable inflation and full employment, can be used to distinguish temporary deviations from the underlying trend in actual GDP growth. This chapter provides the theoretical foundation for potential growth, develops a simple framework to estimate it from readily available macroeconomic indicators, and applies the framework to a sample of developing Asian economies.
Conceptualizing potential growth
Standard economic theory assumes that transitory shocks to the actual growth rate do not significantly affect the dynamics of the underlying trend or the potential growth rate. Several studies have recently challenged this view, however, and have proposed arguments and provided empirical evidence that support the hypothesis that there is a significant relationship between the short-run cyclical behavior of economic growth and long-term performance (e.g., Stiglitz 1993, Cerra and Saxena 2008). This suggests that the GFC may indeed have a significant impact on the future growth trajectory of individual economies.
The possibility of a permanent downward shift affects prospects for sustained increases in output. This is an important medium-term concern for policy makers, particularly in less developed countries where intense pressure exists to deliver sustained increases in living standards. Rodrik (2009), for example, argued that the world after the 2008–2009 crisis would be significantly different from its pre-crisis incarnation, in particular as a milieu for East Asian economies thriving as they did in the 20th century. This is because the crisis has
permanently reduced growth in both productivity and the labor force.1
Likewise, Pritchett and Summers (2014) argued that the days of fast growth in Asia might be numbered because regression to the mean is empirically the most robust feature of economic growth. The statistical analysis showed that rapid growth episodes in developing countries are
frequently followed by significant slowdowns. Moreover, the historical distribution of growth has an average of 2% with a standard deviation that is also 2%.
One approach to addressing the main issues is to examine the supply side of the economy through the concept of potential or natural output growth. The notion of potential output was initially formulated to quantify the ability of an economy to produce output, i.e., its productive capacity. “Potential output” is therefore the highest level of real GDP that can be attained over the long term. By translating levels to growth rates, one obtains the rate of growth of potential output. A limit to output growth exists because of technical, natural, and institutional constraints on the ability to produce.
Growth theory indicates that in the long run economies tend to grow at a rate consistent with the full utilization of productive resources, i.e., the natural or potential growth rate. Short-term shocks can induce temporary deviations from potential growth rate, which give rise to changes in unemployment and inflation, but over time these changes are corrected by price adjustments and growth returns to its potential. What this means is that the high growth rates enjoyed by Asian economies reflected their high potential growth rates. Nevertheless, the precise meaning of this term has evolved over time (Box 2.1.1).
The approach followed in this chapter is consistent with Harrod’s definition of potential growth—the sum of the growth rates of labor productivity and of the labor force—and with Okun’s concept of potential output (Harrod 1939, Okun 1962). Hence, potential output growth is defined as the maximum rate of growth that an economy can achieve consistent with macroeconomic stability, in which there are neither inflationary nor deflationary pressures.
There is broad agreement that Asia’s spectacular growth in recent decades was enabled by its rapid structural transformation, i.e., its capacity to shift resources out of agriculture into sectors of higher productivity (ADB 2013). The speed of an economy’s structural transformation depends largely on institutional barriers that affect, among other things, the reallocation of factors of production, e.g., restrictive labor laws. Given that these barriers are pervasive in many Asian countries and that agriculture is still a large employer (employing 50% of all workers in India and about 30% in the PRC), Asia’s prospects to reengineer high growth will depend largely on eliminating these barriers. Arguably, a significant share of the high growth registered by the Republic of Korea and Taipei,China from the mid-1960s to the mid-1990s, and in the PRC after 1980, is accounted for by the weakening of these barriers, which facilitated the absorption of labor by the manufacturing and service sectors. But most Asian economies still have significant barriers that prevent the development of modern manufacturing and services. Moreover, there are important obstacles that make it difficult to efficiently reallocate factors across firms.
This leads to the important point that potential growth consistent with stable inflation as used in this chapter is conditional on the economy’s institutions and economic structure. Suppose that one could measure additional growth over and above this potential growth that could be achieved with the removal of all institutional barriers
2.1.1 Evolving notions of potential output growth
The idea of potential output has evolved historically. There are four main stages in the progression.
First stage. When Harrod (1939) discussed the “natural growth rate,” this was probably the first formal reference in the literature to the idea of an economy’s full-employment growth rate. Harrod defined an economy’s natural growth rate (ˆgN) as the sum of the growth rates
of labor productivity (ˆyN) and of the labor force (ˆnN), i.e., ˆgN = ˆyN + ˆnN. It represents the maximum sustainable rate
of growth that technical conditions make available to the economic system. At any particular time, actual growth can diverge from the natural growth rate because of various restrictions, rigidities, and constraints, as well as the effect of positive and negative transitory shocks. Nonetheless, actual growth cannot persistently exceed ˆgN as this would
eventually create inflationary pressures, including an excessively tight labor market. With wages rising relative to the price of capital, the economy would adopt more capital-intensive techniques, unemployment would rise again to a comfortable approximation of full employment, and growth would converge on the natural rate. On the other hand, if actual growth were consistently lower than the natural growth rate, the resulting rise in unemployment would trigger the opposite price adjustment process, and falling wages would in due course restore higher employment through the adoption of more labor-intensive production techniques until equilibrium in the labor market was achieved, and actual and natural growth rates were brought into line.
Second stage. It was not until the 1960s that these ideas gave rise to an active body of empirical work around the idea of potential output. Since then, there have been essentially two concepts of potential output in the literature. The first is based on Okun (1962), and the second on Friedman (1968). Okun was interested in the question how much output the economy could produce under conditions of full employment and answered it by introducing the concept of potential gross national product. Okun’s work harked back to Harrod but also accommodated Keynes and the Phillips curve. He argued that potential output is a supply concept of the level of output under the full utilization of factor inputs. However, Okun stressed that the social target of maximum production and employment is constrained by a social desire for price stability. This did not imply, however, that inflation had to be stabilized at a low rate. Okun’s aim was to specify the appropriate fiscal policy
to maximize employment subject to the constraint that inflation should not be excessive. The implied relationship— referred to as Okun’s Law—had practical policy implications under conditions of a stable trade-off between inflation and unemployment (the Phillips curve) and the direct responsiveness of employment to demand (as explored by Keynes).
Third stage. Okun was seriously questioned in the 1970s. The reason was not his concept of potential output. Rather, the rise of inflation during that decade caused disappointment with full employment policies. Friedman (1968) and Phelps (1967) questioned the stability of the inflation–unemployment trade-off embodied in the Phillips curve. They argued that the rate of wage increases was stable at only one rate of unemployment, which was termed its “natural rate” or the “nonaccelerating inflation rate of unemployment.” If unemployment were beneath this natural rate, expectations of rising wages would cause inflation to accelerate without limit. The Friedman–Phelps critique of the Phillips curve had clear significance for the labor market and cast doubt on the feasibility of full employment, which was the touchstone of Okun’s work.
Fourth stage. The result was a radical shift in thinking in the 1980s and 1990s about the relationship between demand pressure and inflation. The outcome was Okun’s concept of the output gap being superseded by variants of Friedman’s natural rate hypothesis. The central bank reaction function presented in Taylor (1993), for example, became a rule observed by many central banks when setting interest rates. This work led by the end of the 1990s and early 2000s to the so-called New Consensus Macroeconomics, also referred to as the New Keynesian Perspective. Essentially, this is a three-equation model consisting of (i) the relationship between interest rates and the output gap, (ii) the Phillips curve relationship between inflation and the output gap, and (iii) the so-called Taylor Rule, which sets interest rates according to the output gap and the difference between actual and targeted inflation. In these models, prices or real wages adjust toward their long-run equilibrium values but do so slowly, with the result that actual output may deviate from the short-term measure of potential output. Consequently, the output gap measures the deviation of actual from potential output that arises as a result of rigidities in prices and wages that prevent them from responding freely to changes in demand and supply.
and distortions in the economy: market imperfections, collusion, rent-seeking, externalities, and government policies that obstruct the mobility of factors, both capital and labor. This would then be the true maximum rate at which the economy can grow. This concept is referred to as the economy’s frontier potential growth. It cannot be quantified, but the idea is clear and useful for developing economies.
2.1.2 Sample economies Azerbaijan Bangladesh Cambodia People’s Republic of China Fiji
Hong Kong, China India Indonesia Kazakhstan Republic of Korea Malaysia Pakistan
Papua New Guinea Philippines Singapore Sri Lanka Taipei,China Tajikistan Thailand Turkmenistan Uzbekistan Viet Nam
While the difference between potential growth and frontier potential growth is not large in advanced economies, the distinction is important for economies riddled with inefficiencies. Policies and reforms that ease some of these constraints in the latter group can have salutary effects on potential growth, especially if it languishes far from the frontier. Understanding these constraints can help policy makers propel their economies closer to their true limit.
Estimating potential growth
Potential output is not directly observed. As such, the first step in analyzing developing Asia’s potential is to develop an empirical model. This section develops a simple empirical model and applies it to a sample of 22 developing economies in Asia (Box 2.1.2). While limited data availability constrained selection for the sample, the group includes the three
largest economies—the PRC, India, and Indonesia—and representatives of all subregions in developing Asia. As the sample accounts for more than 98% of GDP in the region in 2014, the aggregate results should broadly represent the region’s experience.
At present, most economists view economic growth as the sum of a cyclical component and a permanent component. The former captures the business cycle or demand
fluctuations and the latter the long-run trajectory of growth.
This is the “natural” or potential growth rate of the economy, consistent with the full utilization of productive resources—in particular full employment of the labor force accompanied by stable inflation. There are various methods to estimate the potential growth rate, which are summarized in Box 2.1.3.
Of the methods listed in the box, this chapter uses a multivariate estimate that relies on information on GDP growth and inflation. This is obviously an improvement over the univariate filters. Unlike the growth accounting approach, it does not require a series of capital stock, which is lacking in most Asian economies. And, unlike the output identity approach, the method employed in this chapter provides an explicit link with concepts of macroeconomic stability. Before the model is estimated, univariate filters are applied to the data series to extract the underlying trends of each variable.
As noted above, deviations of actual growth from the potential rate cause resources to be either overutilized or underutilized, giving rise to changes in the unemployment rate and, consequently, to inflationary or deflationary pressures. Consequently, potential output growth is associated with a stable inflation environment. The framework used in this chapter to estimate the potential growth rate of the Asian economies builds on this intuition.
Potential growth as modeled in this chapter is consistent with the Harrod (1939) concept of the natural rate of growth. The natural rate of growth is described as both the potential growth rate toward which the economy tends in the long-run and the short-term upward limit to growth, which turns cyclical expansions into recessions.
Since by definition it is the sum of the growth rates of labor productivity and the labor force, potential growth can be usefully decomposed into these two elements once it is estimated. Note that when actual growth is equal to potential, employment grows at the same rate as the labor force, keeping the unemployment rate constant. As such, when actual growth is consistently slower than the natural rate, unemployment will rise, and vice versa.
As a consequence, a convenient method to estimate the natural growth rate relies on the relationship between unemployment and output growth formalized in Okun’s Law (e.g., León-Ledesma and Thirlwall 2002). Such a choice is complicated, however, by the lack or unreliability of labor market data for many Asian economies. This problem is addressed by noting that a natural extension of the concept of the natural growth rate is the link with the relationship between unemployment and inflation typified in the Phillips curve. The process, as described in Box 2.1.4, is to estimate the relationship of the gap between actual and expected inflation to the gap between actual growth and its natural rate. The model that relates output growth and inflation generates annual estimates of the potential growth rate.
Potential growth trends in developing Asia
When applied to developing economies with significant surplus labor, the model suggests that they can grow fast until the surplus is eliminated. Economies with surplus labor have relatively high labor force growth rates, many unemployed and underemployed workers, and low wages. Under these circumstances, there is room to grow
2.1.3 Methods to estimate potential growth
Potential growth, unlike actual growth, is not directly observable and hence has to be estimated. Some of the most widely used methods are as follows:
Univariate filters. These are statistical procedures that are not based on any underlying economic theory and use information only on GDP. Their objective is to statistically remove the cyclical component of a series from the raw data. The most widely used is the Hodrick–Prescott, but some other filters are the Baxter–King, Christiano– Fitzgerald, Beveridge–Nelson, and Corbae–Ouliaris.
Multivariate estimates. These approaches use information from several economic series to obtain estimates of potential output. The models are typically based on a structural theory. For example, they estimate the rate of economic growth consistent with “macroeconomic stability.” One simple model is a bivariate structural time series linking temporary fluctuations in output to inflation, and postulating that the output gap is positively related to inflation pressures. An extended version could add an Okun’s Law equation relating unemployment and output growth.
Growth accounting approach. This is also a multivariate approach in that it uses information on variables related through an aggregate production function: output, employment, capital, and residually measured total factor productivity. Typically, authors derive a decomposition of the sources of growth using a Cobb–Douglas production function such as (1 )
( ) t t t t t
Y =A Kα L H −α. In most cases it is not estimated econometrically; rather, factor markets are assumed to be competitive, so the labor and capital elasticities equal the factor shares in national income, and these are imposed to derive total factor productivity growth.
Output identity. This approach decomposes output (Y) multiplicatively as the product of a series of terms, e.g., labor productivity in hours (Y/H), hours per employee (H/L), the employment rate (L/P), and working-age population (P), such that
Y * H * L *
H L P
Y P. These series can also be filtered to calculate their trend, which is then interpreted as the potential level. Like the production function approach, this method does not explicitly link the estimation of trend growth to the estimation of the output gap and inflation.
2.1.4 Multivariate estimates of potential growth
The model used for estimation has three underlying pillars: (i) Harrod’s notion of the natural growth rate, (ii) Okun’s Law relating unemployment to output, and (iii) the Phillips’ curve relating inflation to output.
Since the natural growth rate is defined as the sum of the growth rates of labor productivity and the labor force, unemployment will rise whenever the actual rate of growth (ˆgt) falls below the natural rate ˆ( N)
t
g , and it will fall when
ˆgt rises above ˆN t
g . This yields the following specification of Okun’s Law:
ˆ
t t
U σ ςg
Δ = − (1)
where 'Ut is the percentage change in the unemployment rate and the natural growth rate given by (V9). This specification and its variants have been widely used in the literature to estimate ˆN
t
g for countries and regions and also to investigate its possible endogeneity (e.g., León-Ledesma and Thirlwall 2002, Lanzafame 2010). The specification in equation (1) presents two issues that are addressed in this chapter. First, the model produces only a single estimate of the potential growth rate for the time period under analysis. Since its evolution over time is what is important, a time-varying parameter approach is applied to estimate a time series for ˆN
t
g . Second, the unemployment rate and, more generally, labor market data are notoriously unreliable for some of the economies in the sample.
To deal with this data problem, Harrod’s definition of ˆN
t
g is linked to the relationship between unemployment and growth. The potential growth rates of the Asian economies are then estimated based on an aggregate supply model. In the long run, unemployment will be constant when it is equal to the nonaccelerating inflation rate of unemployment (NAIRU). Therefore, the potential growth rate can be defined as that growth rate consistent with N
t t
U =U , which implies 'Ut = 0. Okun’s relation in terms of the NAIRU can be rewritten as follows: ˆ ˆ ( ) N N t t t t t U =U −β g −g (2) where the Okun coefficient (Et) and the NAIRU ( )
N t
U are assumed to vary over time.
The relationship between inflation and unemployment is given by the following Phillips curve:
( ) e N t t t Ut Ut S S J (3) where St and e t
S are, respectively, the actual and expected inflation rates, while Jt is a time-varying parameter. By substituting (3) into (2), the following equation is obtained: ˆ ˆ ( ) e N t t t gt gt S S I (4)
where I E Jt t t. The specification in (4) formalizes an aggregate supply model with time-varying parameters.
To estimate the model in (4), an estimate of the expected inflation rate e
t
S is required. Because of limited data for expected inflation, e
t
S is modeled as a function of the actual inflation rate St with two possible specifications. One specification, expected inflation in time t, is a time-varying function of actual inflation in t plus a random shock:
e
t t t t
S D S H (5)
The estimated model in this case is as follows: (1 ) ˆ ˆN t t t t t t g g D S H I (6)
The second specification assumes an extreme form of adaptive expectations (a random walk), with expected inflation in t equal to actual inflation in t–1 plus a random shock:
1
e
t t t
S S (7)H
The second model is as follows: 1 ˆ ˆN t t t t t g g S H I ' (8) The constant term ˆN
t
g in equations (6) and (8) provides
an estimate of potential growth over time. To take account of the possible effects of the degree of openness on the slope of the Phillips curve, equations (6) and (8) are augmented with the share of imports in GDP. These equations are specified in state-space form, and the Kalman filter is used for estimation. This is a statistical procedure that can produce time-varying coefficients (Harvey 1989).
faster without stoking inflationary pressures. When surplus labor is eliminated, wages and possibly inflation start increasing. That is when growth is at its potential. It is true that a country with surplus labor can have rising inflation, but the likely explanation in such a situation is that constraints are creating bottlenecks that hamper the reallocation of labor. This implies that an economy’s potential growth rate could be higher if such bottlenecks were eased or eliminated.
Meanwhile, in the case of the Central Asian economies, their dependence on natural resources can affect the estimates through the effect of commodity price shocks on domestic inflation. However, it is
difficult to state a priori the sign of this relation, which possibly should be stronger in more open economies. This can be controlled for, if only indirectly, by including in the model the share of imports in GDP (Romer 1993). The problem is the lack of sufficiently long time series, with the consequence that meaningful results are difficult to obtain. The model being estimated is therefore based on certain assumptions that may not exactly fit oil-dependent economies. Box 2.1.5 discusses the how well the model performs with the given sample.
Graphs of estimated potential growth rates and actual growth rates are presented in Figure 2.1.2. These are for 12 of the 22 developing
economies in Asia under consideration.2 The corresponding graphs for
the other 10 economies are given in the Annex.
These graphs show that the potential growth rate was more stable than actual growth, and that it was fairly high and/or increasing in the 1980s and 1990s in most economies. The pattern aligns with expectations. The graphs also suggest that, in most cases, the estimated potential growth rate was higher in 2000 than in 2014. Moreover, the trend was either stable or declining during 2008–2014. Figure 2.1.3 shows the difference between the period average rates for potential growth and actual growth between 2000–2007 (pre-crisis) and 2008–2014 (post-crisis).
It is worth noting that during 2008–2014 potential growth declined in Asia’s advanced economies as well as in some of the major economies, including the PRC and the Republic of Korea. Thailand suffered a significant decline in its potential growth rate, which is the lowest in Southeast Asia. Potential growth increased, however, in Indonesia, Pakistan, the Philippines, and Uzbekistan. Meanwhile, Bangladesh, Fiji, and India maintained the same pace in both subperiods. Overall, comparing pre-crisis and post-crisis periods, the decline in potential
growth accounts for 39.6% of the decline in actual growth.3 This implies
that about 60% of the decline appears to be a temporary effect of the business cycle.
2.1.5 How well does the aggregate supply model of potential growth perform?
To gauge whether the model produces sensible estimates of potential output growth, test its key insight: Inflationary pressures should arise when the gap between actual and potential growth rates widens. The test is carried by pooling all data. The following regression is estimated using a fixed effects model: ˆ ˆ ( N) it i git git it S - Z H '
where ' is the change in the inflation rate and ˆSit ( ˆ ) N it it
g g
is the difference between actual and potential growth rates. The expectation is that Z > 0 and statistically significant. The model is first estimated for a large sample of 71 economies across the world, and then for the sample of 22
developing Asian economies plus Japan. Results confirm the hypothesis that, when the gap between actual and potential growth widens, inflationary pressures emerge. In the larger sample, results confirm that, where inflation rates are below 25%, for each percentage point of actual growth in excess of the natural growth rate, the inflation rate increases by about 0.12 percentage points. Where inflation is above 25%, the relationship breaks down. Meanwhile, in the sample of Asian economies, results show that, where inflation rates are below 45%, for each percentage point of actual growth in excess of the natural growth rate, the inflation rate increases by about 0.2 percentage points.
2.1.2 Potential growth rate estimates and actual growth rates –6 –3 0 3 6 9 12 1961 1971 1981 1991 2001 2011 0 4 8 12 16 % % % 1987 1992 1997 2002 2007 2012 –10 –5 0 5 10 1960 1970 1980 1990 2000 2010 –5 0 5 10 15 20 1976 1983 1990 1997 2004 2011 –4 0 4 8 12 16 1971 1979 1987 1995 2003 2011 –15 –10 –5 0 5 10 15 1994 1998 2002 2006 2010 2014 India Indonesia People’s Republic of China
–20 –10 0 10 20 1971 1979 1987 1995 2003 2011 % % % %
Singapore Taipei,China Thailand Philippines 0 4 8 12 16 1999 2004 2009 2014 Kazakhstan % –10 0 10 20 1968 1976 1984 1992 2000 2008 Republic of Korea % –10 –5 0 5 10 15 1961 1971 1981 1991 2001 2011 % Malaysia –3 0 3 6 9 1985 1990 1995 2000 2005 2010
% Pakistan Papua New Guinea
% –2 0 2 4 6 8 10 12 2002 2005 2008 2011
Note: Broken line refers to potential growth. Solid line corresponds to actual growth. Source: ADB estimates.
Key features of potential growth trends in the four largest economies in the sample, which accounted for about 80% of developing Asia’s actual GDP in 2014, were as follows:
People’s Republic of China. During the expansionary phase up to 2007, before the GFC, the PRC economy operated at potential or slightly above it. From 2008, a gradual decline in potential growth is observed to be about 7.91% in 2014, just slightly above actual growth at 7.4%. Post-crisis average potential growth declined by 1.11 percentage points relative to the pre-crisis average. During 2008–2014, inflation slowed as the deviation of actual growth from potential growth was only slightly negative. This is consistent with the theoretical framework underlying the model, which states that a decline in the inflation rate is associated with the actual growth rate
being below potential.4
India. For most of the estimation period, India’s potential
growth rate was quite stable.5 A substantial increase starting
in the early 2000s is observed, with potential growth peaking in 2007 at 8.6%. As with the PRC, potential growth declined in the aftermath of the GFC, falling to 6.29% in 2014.
However, the average in 2008–2014 was only 0.06 percentage points lower than in 2000–2007. The average deviation of actual from potential growth was slightly negative. Also as with the PRC, this indicates that India’s economy grew at close to its potential during this period.
Republic of Korea. The economy’s potential growth rate has been declining for quite some time. It experienced a significant decline during the Asian financial crisis of 1997– 1998, after which it recovered slightly. However, the broader trend since 2001 has been negative. The 2008–2014 average was 2.09 percentage points lower than in 2000–2007. The 2008–2014 averages of actual and potential growth were also very close to each other, the deviation being slightly negative. This is consistent with a slightly declining inflation rate. In 2014, the Republic of Korea’s potential growth rate was estimated at 3.33%, very close to actual growth.
Indonesia. For most of the estimation period, Indonesia’s potential growth rate was stable. However, its potential growth rate declined during the Asian financial crisis of 1997–1998 to 3% from a peak of 8.4% in 1995. It then recovered to an average of almost 5% in 2000–2007 and increased to 5.81% during 2008–2014. During this period, actual and potential growth rates were close, so that the average change in inflation was stable. Among developing
Asia’s four largest economies, only Indonesia registered a significant increase in potential growth during 2008–2014, rising 0.86 percentage points higher than the 2000–2007 average.
In Figure 2.1.4, the weighted average of the estimated potential growth rates of 13 Asian economies (Asia-13) for which consistent estimates since 1988 are available is calculated. These economies— Bangladesh, Fiji, India, Indonesia, the Republic of Korea, Malaysia,
2.1.3 Differences in average actual and potential growth rates, 2000–2007 versus 2008–2014 –15 –10 –5 0 5 10 Azerbaijan Turkmenistan Rep. of Korea Kazakhstan Cambodia Hong Kong, China Tajikistan Japan Taipei,China Viet Nam PRC Sri Lanka Thailand Singapore Malaysia India Fiji Bangladesh Philippines Indonesia Pakistan Uzbekistan Papua New Guinea
Potential Actual
Notes: For Fiji, pre-crisis growth rates are for 2001–2007. For Papua New Guinea, post-crisis growth rates are for 2008–2013. Singapore and Thailand used a model augmented with financial factors (Felipe et al. 2015). For Tajikistan, pre-crisis actual growth is for 2001–2007, and potential growth is for 2002–2007.
Pakistan, the Philippines, the PRC, Singapore, Sri Lanka, Taipei,China, and Hong Kong, China—represent almost 94% of developing Asia’s actual GDP in both 1988 and 2014. The weighted average provides a more aggregate picture of these growth dynamics and will help answer the question whether there is evidence that Asia is entering a new normal of lower potential growth. Average potential growth for two other subgroups is also plotted: One is the same economies but excluding the PRC (Asia-12) and the other includes only the Republic of Korea, Singapore, Taipei,China, and Hong Kong, China (Asia-4). A comparison of Asia-13 and Asia-12 averages allows the assessment of the influence of the PRC on the potential growth performance of the region as a whole.
In the case of Asia-4, potential growth declined from an average of about 9% in 1988 to just over 3% in 2014. The aggregate behavior of Asia-13 provides a more complex and nuanced view of Asia’s potential growth. First, potential growth increased prior to the Asian financial crisis of 1997–1998. The series peaked at a range of 7.3%–7.4% in 1994–1996. Potential growth declined during the crisis and bottomed out in 1998 at 5.36%. It then recovered and increased from 1999 to 2007, peaking at 8.45%. After the GFC in 2008, potential growth started declining, falling to 6.68% in 2014, almost 2 percentage points below the peak. The significant decline from the peak indicates that the region may have entered a new normal of lower potential growth, an issue that will be explored further in the third section of this theme chapter.
It is also important to note that developing Asia’s dynamics of potential growth are increasingly determined by the growth performance of the PRC. The PRC share in regional GDP increased to almost 55% in 2014 from only 25% in 1988. In this context, the surge in regional potential growth between 1998 and 2007 was mostly the result of the phenomenal growth performance by the PRC (Figure 2.1.5). The consequence was that the PRC contributed about three-quarters of the overall rise in potential output growth during this period. Specifically, the PRC contributed 2.31 percentage points out of the total increase of 3.09 percentage points. Of that 2.31 percentage points, 1.09 is derived from its rising potential growth rate, and 1.22 can be attributed to the growing PRC share in developing Asia’s GDP. Most of the remaining increase in potential growth was contributed by India and the Republic of Korea, while several economies made negative contributions to the regional average.
In the same vein, the decline from the onset of the GFC to 2014 was determined primarily by events in the PRC, India, and the Republic of Korea. However, the latter two economies played a bigger role in this episode. Specifically, about one-third of the 1.78 percentage point fall in average
2.1.5 Contributions to change in potential growth
–20 0 20 40 60 80 Percentage share Sri Lanka Fiji Pakistan Philippines Bangladesh Indonesia Malaysia Singapore Taipei,China Hong Kong, China People’s Republic of China India Republic of Korea Between 2008 and 2014 Between 1998 and 2007
Source: ADB estimates.
2.1.4 Estimates of average potential growth rate
% 0 2 4 6 8 10 1988 1993 1998 2003 2008 2013 7.39% (Ave. 2000–07) 8.45% 6.68% 7.07% (Ave. 2008–14) Asian-4 average
Asian-12 average (excluding the PRC) Asian-13 average
Asia-4 = Republic of Korea, Singapore, Taipei,China, and Hong Kong, China; Asia-12 = Asian-13 minus the People’s Republic of China; Asia-13 = Bangladesh, the People’s Republic of China, Fiji, India, Indonesia, the Republic of Korea, Malaysia, Pakistan, the Philippines, Singapore, Sri Lanka, Taipei,China, and Hong Kong, China.
Notes: Regional potential growth rate is the average of the individual economies’ growth rates, weighted by their share in actual GDP. The regional average including the 22 economies that constitute developing Asia (i.e., the 22 in the study, excluding Japan) since 2000 is as follows (Tajikistan is not included in 2000 and 2001. Its weight is only 0.3%): 2000: 7.04%; 2001: 6.18%; 2002: 6.65%; 2003: 6.75%; 2004: 7.40%; 2005: 7.69%; 2006: 8.40%; 2007: 8.38%. 2008: 7.21%; 2009: 6.87%; 2010: 7.68%; 2011: 7.00%; 2012: 6.70%; 2013: 6.71%; 2014: 6.56%.
potential growth can be attributed to the Republic of Korea, which experienced a large fall in both potential growth and GDP share— nearly 7 percentage points—from 2007 to 2014. India’s GDP share and potential growth rate also declined, by 2 and 2.3 percentage points, respectively. As a result, its contribution to the aggregate decline was also approximately a third. Meanwhile, the decline in potential growth in the PRC was offset by the rise in its regional share of GDP, so that its net contribution to the regional decline was only 0.24 percentage points, or 13% of the total.
The picture without the PRC (Asia-12) is very similar to that of the Asian-13 but, as expected, potential growth without the PRC is lower both before and after the GFC. As with Asia-13, the decline during 2008–2014 was significant. The substantial decline in Asia-12 average potential growth with respect to the earlier part of the 1990s is also noted. Another common feature is that Asia-12 average potential growth during the Asian financial crisis was lower than during the GFC. With or without the PRC, the behavior of potential growth clearly shows a turning point after the GFC.
The question remains whether this slowdown in potential growth is a new normal for the region—and what, if anything, policy makers can do to counteract this trend. Future gains in living standards and poverty elimination rest on a grasp of potential growth. Identifying the institutional and policy constraints that keep an economy from achieving its frontier potential growth will be critical to counteract the current slowdown. Knowing which factors have driven potential growth in the past can guide policy makers’ efforts to meet today’s growth challenge.
Determinants of potential growth
Several approaches are used to tease out what is behind the recent slowdown in potential growth. First, the time-series estimates of potential growth for the 22 developing economies in Asia can be decomposed using a set of simple techniques that stem from the definitions of the key concepts. From the Harrod (1939) concept of the natural growth rate, potential growth is the sum of labor force growth and potential labor productivity growth—that is, growth in potential output per worker. Potential labor productivity growth itself can be decomposed to show it driven by such factors as within-sector productivity growth, the effect of employment reallocation across sectors, capital accumulation, and total factor productivity growth.
Second, an econometric model is estimated using cross-country data for a larger sample of countries including those outside of Asia to establish the determinants of potential growth. The objective is to obtain robust and reliable estimates of variables significantly correlated with the potential growth rate.
Finally, firm-level data are used to study the role of institutional obstacles in generating distortions that cause resource misallocation. This exercise can help shed light on policies that could help shift an economy closer to its frontier potential growth.
Potential labor productivity growth
Following Harrod’s definition of the natural growth rate, potential labor productivity growth is potential growth less the growth rate of the labor force. Because labor market data in many developing economies in Asia are unreliable, working-age population, which is the population aged 15–64, is used as a proxy for the labor force. These data are filtered using appropriate statistical techniques to purge short-term variability, which is caused largely by transitory migration flows.
Figure 2.2.1 shows both potential and actual labor productivity growth rates corresponding to 12 Asian economies whose potential growth was graphed in the previous section. The PRC displays a high potential labor productivity growth rate—significantly higher than that of India. This explains a large part of the difference in potential output growth between these two economies. Consistent with their transition into high-income economies, the Republic of Korea and Taipei,China show marked declines in potential labor productivity growth. This is less obvious in Singapore. Actual and potential labor productivity growth tend to track each other quite closely in most economies. However, the Philippines is a notable exception, with actual labor productivity growth lagging potential for most of the period.
Figure 2.2.2 reports the breakdown of potential output growth into potential labor productivity growth and filtered labor force growth for the periods 2000–2007 and 2008–2013. With a few exceptions— Malaysia, Pakistan, and Singapore—potential labor productivity growth
2.2.1 Potential and actual labor productivity growth rates
India Indonesia People’s Republic of China
Singapore Taipei,China Thailand Philippines % –5 0 5 10 15 1990 1995 2000 2005 2010 % –20 –10 0 10 20 1990 1995 2000 2005 2010 % –5 0 5 10 1990 1995 2000 2005 2010 % –20 –10 0 10 20 1990 1995 2000 2005 2010 % –10 –5 0 5 10 15 1990 1995 2000 2005 2010 % –2 0 2 4 6 8 10 1990 1995 2000 2005 2010 % –10 –5 0 5 10 15 1990 1995 2000 2005 2010 Kazakhstan % –40 –20 0 20 1990 1995 2000 2005 2010 Republic of Korea % 0 2 4 6 8 10 1990 1995 2000 2005 2010 Malaysia % –10 –5 0 5 10 1990 1995 2000 2005 2010 Pakistan % –10 –5 0 5 10 1990 1995 2000 2005 2010
Papua New Guinea
% –2 0 2 4 6 8 10 12 2002 2005 2008 2011
Note: Broken line refers to potential growth. Solid line corresponds to actual growth. Source: ADB estimates.
is the main contributor to potential output growth. On average, potential labor productivity growth accounted for 78% of potential output growth during 2000–2007 and 86% during 2008–2013.
Meanwhile, Figure 2.2.3 shows the change in annual average potential output growth between the periods 2000–2007 and 2008–2013. There is a decline in most economies. The figure also shows in each bar the change in the two components. In the PRC, potential output growth declined by about 1 percentage point between the two subperiods. Most of the decline is accounted for by the slowing of working-age population growth, with potential labor productivity growth actually increasing slightly. On the other hand, India saw a very small increase in potential output growth between the two periods. This was accounted for by the increase in potential labor productivity growth, which was slightly larger than the decline in working-age population growth. In the Republic of Korea, most of the decline in potential growth is accounted for by the decline in potential labor productivity growth. And, in Indonesia, most of the increase in potential output growth was accounted for by the increase in potential labor productivity growth.
Decomposition of potential labor
productivity growth
Using shift-share analysis, the estimates of labor productivity growth are decomposed into productivity growth within the sector, changes in the employment structure, slack in the economy, capital deepening, and total factor productivity growth (Foster-McGregor and Verspagen 2016). This provides a simple but very useful way to assess which component of potential labor productivity growth is the largest and matters the most in explaining labor productivity growth. Box 2.2.1 shows how these decompositions are interpreted.
Equation (3) in Box 2.2.1 shows that potential labor productivity growth can be decomposed into the sum of the within, static, and dynamic effects, plus two terms that capture an output gap and a labor gap. This decomposition allows the comparison of the changes in the two gaps with the three structural effects. Results are presented in Figure 2.2.4 along with the average potential labor productivity growth for 1990–2011.
In most economies, the contribution of the combined gaps to potential labor productivity growth is relatively small compared to those of the structural effects. Kazakhstan, the Philippines, and Sri Lanka are the main exceptions, all with gaps making positive contributions. The positive contribution means that potential labor productivity growth has been faster than actual labor productivity growth.
2.2.2 Contributions to potential output growth
MAL PHI SIN KOR TAP THA VIE PAK KAZ SRI AZE BAN FIJ PNG CAM UZB TAJ % Potential labor productivity growth (2000–2007) Trend labor force growth (2000–2007) Potential labor productivity growth (2008–2013) Trend labor force growth (2008–2013) PRC HKG IND INO 20 0 40 60 80 100
Note: For Fiji and Papua New Guinea, first period refers to 2002–2007. For Taipei,China, Cambodia, and Uzbekistan, second period refers to 2008–2012. For Tajikistan, first period refers to 2002–2007 while second period refers to 2008–2012.
Source: ADB estimates.
2.2.3 Change in potential output growth and its components between 2000–2007 and 2008–2013
% Potential labor productivity growth Trend labor force growth Annual potential output growth PRC HKG IND INO MAL PHI SIN KOR TAP THA VIE PAK KAZ SRI AZE BAN FIJ PNG CAM UZB TAJ –12.2 –12.3 –6 –4 –2 0 2 4 6
Note: For Fiji and Papua New Guinea, first period refers to 2002–2007. For Taipei,China, Cambodia, and Uzbekistan, second period refers to 2008–2012. For Tajikistan, first period refers to 2002–2007 while second period refers to 2008–2012.
2.2.1 Decomposition of labor productivity growth, potential labor productivity growth, and total factor productivity growth
The procedure for decomposing potential labor productivity growth can be depicted in stages:
1. Decomposition of actual labor productivity growth. A well-known structural decomposition of labor productivity growth is given by the following equation (omitting the subscript ‘t’ for time):
1 0 1 0 1 0 1 0 0 0 0 0 0 ( ) ( ) ( )( ) ˆ ˆ ˆ i i i i i i i i i i i i i y y s s y y s s g l y S y y y y { { ¦ ¦ ¦ (1)
where ˆg is the growth rate of GDP, ˆl is the growth rate of aggregate labor input, y is labor productivity and ˆy denotes actual labor productivity growth, and si denotes the share of industry i in aggregate employment. The three terms on the right hand side are the within effect, or the contribution of productivity growth rates within each sector i; the static
effect, or the productivity effect of relocating labor that
results from differences in productivity levels at the start of the period; and the dynamic effect, or the productivity effect that results from relocating labor from one sector to another, at the same time taking into account the change in the productivity growth rate over the period. The sum of the last two is equal to the effect of employment reallocation across sectors. The symbol 6 denotes summation with this operation carried out for each of the three effects across the nine sectors for which data are available: agriculture, mining, manufacturing, construction, utilities, trade, transport and commerce, the public sector, and finance, insurance, and real estate.
2. Potential labor productivity growth and the output and labor gaps. The same structural decomposition of labor productivity growth is applied to the potential rate of labor productivity growth (ˆyN) by rewriting the definition of
potential growth rate (gN) as follows:
ˆ ˆ
ˆ ˆ ˆ ˆ
ˆN N ˆN ( ) ( N) ( ˆN),
y {g n { g l gg l n (2)
where ˆnN is the growth rate of potential labor force,
ˆ ˆ ˆ
(g l { is the growth rate of actual labor productivity, ) y
ˆ ˆ (ggN)
is the difference between the actual rate of output growth and the potential rate (or the output gap), and ˆ(lnˆN)
is the difference between actual employment growth and the growth rate of the potential labor force (and equal to the growth rate of the participation rate). This is called the labor gap. These last two terms reflect slack in an economy that is not operating at its potential.
Substituting equation (1) into equation (2), yields a decomposition of potential labor productivity growth:
1 0 1 0 0 0 0 0 1 0 1 0 0 ( ) ( ) ˆ ( )( ) ˆ ˆ ˆ ˆ ( ) ( ) N i i i i i i i i N N i i i i i y y s s y S y y y y y s s g g l n y { ¦ ¦ ¦ (3)
The first three terms have the same interpretation as in equation (1), while the last two are the output and labor gaps. A negative output gap and a positive labor gap add to potential labor productivity growth.
3. Total factor productivity growth. Total factor productivity (TFP) is often used as a measure of the efficiency with which both labor and capital are used and is sometimes preferred to labor productivity. The growth rate of TFP can be decomposed into the same three terms as labor productivity growth by replacing labor productivity in equation (1) with TFP (I). The resulting decomposition is as follows: 1 0 1 0 1 0 1 0 0 0 0 0 0 ( ) ( ) ( )( ) ˆ i i i i i i i i i i i i i s s s s S I I I I I I I I I { ¦ ¦ ¦ (4)
The same three terms representing the within, static, and dynamic contributions to TFP growth ( )Iˆ in this case are present.
4. Extended decomposition of potential labor productivity growth: the role of TFP growth and of capital deepening. Combining TFP growth decomposition with decomposition of potential labor productivity growth provides a more detailed decomposition of the latter that allows an
estimation of the effects of both capital deepening and TFP growth. From a production function that assumes perfectly competitive factor markets, actual labor productivity growth can be written as follows:
ˆ ˆ ˆ ˆ ˆ
(g l { ) I (1 D)(k l ) (5) where ˆk is the growth rate of aggregate capital, (1 – D) is the share of capital in GDP, and ˆI is the growth rate of TFP. This equation states that the growth rate of actual labor productivity is equal to TFP growth plus the weighted growth rate of the ratio of capital over labor, with the weight being the share of capital in GDP. Replacing ˆI in equation (5) with the expression in equation (4) and inserting this into equation (2) results in the following decomposition of potential labor productivity growth:
1 0 1 0 0 0 0 0 1 0 1 0 0 ( ) ( ) ˆ ( )( ) ˆ ˆ ˆ ˆ ˆ ˆ (1 )( ) ( ) ( ) N i i i i i i i i N N i i i i i s s y S s s k l g g l n I I I I I I I D I { ¦ ¦ ¦ (6)
This relates potential labor productivity growth to the within, static, and dynamic contributions to TFP growth, capital deepening, and the output and labor gaps.
The analysis implies that, on average, the within effect accounts for 91% percent of potential labor productivity growth, and the static effect for 15%. Meanwhile, the dynamic effect is small, and the contribution of the two gaps is
negative. It can therefore be concluded that the failure to realize potential because of the two gaps has a minimal role in explaining potential labor productivity growth.
Subsequent decomposition accounts for capital deepening and TFP growth (Figure 2.2.5). As shown in Box 2.2.1, the contribution of TFP growth is equal to the sum of the within, static, and dynamic effects. Interestingly, the contribution of capital deepening is always positive and high in many countries. Hence, a significant portion of overall labor productivity growth comes from capital deepening rather than from pure efficiency improvements. This also implies that actual labor productivity growth is higher than TFP growth in all cases. As in the previous decomposition, the within effect is found to make the dominant contribution to potential labor productivity growth.
Factors affecting structural transformation
The findings show that the structural transformation effects (that is, the static and dynamic effects) contributed much less than the within effect. This raises the question what has prevented the faster reallocation of labor. Felipe et al. (2014) documented that the share of agricultural employment in the Republic of Korea recorded a 1.23 percentage point decline per annum during 1962–2013, falling from 69% to 6%. Meanwhile, the corresponding decline for the PRC was 1 percentage point per annum, from a share of 82% to 31%, and for Taipei,China 0.88 points, from a share of 50% to 5%. These are among the fastest declines in history, even faster than the declines experienced by the advanced economies in the 19th and 20th centuries. Meanwhile, India, Pakistan, and many other economies still have very high agricultural employment shares, which are declining very slowly.
Figure 2.2.6 shows the employment structure of typical low-, middle-, and high-income country. Probably, the most salient feature is the different shares of agriculture: 39% of total employment in low-income economies, 17% in middle-income economies, and 2% in high-middle-income economies. The flip side of this is the share of services.
Structural transformation has been much faster in some Asian economies than in others. Differential paces of economic growth can, by themselves, generate different speeds of structural change that reflect the relative productivity growth of different sectors, or changes in the composition of demand (Herrendorf et al. 2014). Impediments to the reallocation of factors of production are also important. These can be related to government failure or market failure caused by coordination problems and underdeveloped financial markets (Sen 2016).
2.2.4 Structural decomposition of potential labor productivity growth, 1990–2011 % Dynamic effect Static effect Gaps Within effect % Potential labor productivity growth
–4 –2 0 2 4 6 8 10 –40 –20 0 20 40 60 80 100
THA SRI HKG SIN AZE BAN INO MAL KOR TAP IND VIE PRC KAZ PAK PHI
Note: For Kazakhstan, 1999–2008. For Pakistan, 1995–2008. For Thailand, 1994–2011. For Sri Lanka, 1995–2011. For Azerbaijan, 2001–2011. For Bangladesh, 2003–2009. For Viet Nam, 1996–2011. Source: ADB estimates.
2.2.5 Extended decomposition of potential labor productivity growth, 1990–2011 % Gaps Capital deepening TFP growth % 1 2 3 4 5 6 7 8 –40 –20 0 20 40 60 80 100
PAK JAN THA VIE BAN MAL HKG SIN INO PHI TAP KAZ SRI IND KOR PRC AZE
Potential labor productivity growth (average 1990–2011)
PRC = People’s Republic of China, TFP = total factor productivity. Source: ADB estimates.
Government failures that slow down the rate of structural transformation often occur in four policy areas. First, land reform and land-use policies affect the reallocation of labor resources from low- to high-productivity sectors, albeit indirectly. Widespread land reform through the fair redistribution of agricultural land can boost agricultural productivity and provide surplus labor that may be reallocated to industry. Agrarian reform can also effect income redistribution that encourages the formation of human capital and increase aggregate demand. The Republic of Korea and Taipei,China successfully implemented widespread land reform prior to industrialization, allowing their economies to undergo rapid structural transformation from the 1970s to the 1990s. Other economies in Asia attempted to implement land reform but with few tangible results.
Second, labor policies that are geared toward protecting labor welfare impose rigidities in nominal wages and introduce barriers to labor reallocation. These regulations provide incentive to substitute capital for labor and can slow down the pace of investment. As such, labor market regulations potentially impede structural transformation when they are overly rigid. One problem partly responsible for India’s slow pace of structural transformation is its rigid labor market as stipulated in the Industrial Disputes Act, 1947.
A third government failure is migration policy that restricts worker mobility. Limiting the mobility of labor directly affects the availability of this resource and can create geographic imbalance between the availability of qualified workers and demand for them. An example of a restrictive migration policy is the hukou system of residency permits in the PRC, which constrains the free flow of labor between urban and rural areas and consequently slows the pace of structural transformation.
Lastly, governments inadvertently raise the cost of business by imposing product and market regulations. These slow structural transformation through their negative effects on investment and, consequently, firms’ scale of operations. While governments often
2.2.6 Typical employment structures by type of economy
East Asia High income Utilities 1% Construction 7% East Asia Low income East Asia Middle income Agriculture 17% Agriculture 2% Agriculture 39% Mining 1% Mining 0% Manufacturing 20% Manufacturing 16% Manufacturing 15% Utilities 1% Utilities 1% Construction 8% Trade, restaurants and hotels 18% Public and social services 23% Public and social services 31% Public and social services 17% Construction 6% Mining 1% Trade, restaurants and hotels 14% Trade, restaurants and hotels 22% Transport, storage and communication 7% Transport, storage and communication 6% Transport, storage and communication 4% Finance, insurance, real estate and business services
14%
Finance, insurance, real estate and business services
6%
Finance, insurance, real estate and business services
3%
Notes: These are derived from Lowess regressions (Foster-McGregor and Verspagen 2016). The shares in the three charts are the predicted values of these regressions at three distinct values of GDP per capita: $24,189 (the 2011 value for Taipei,China as the high-income reference), $8,737 (the 2011 value for the PRC as the middle-income reference), and $3,372 (the 2010 value for India as the low-income reference). To be precise, the employment structure of these three economies is not used, but instead the predicted values of the employment structure derived from the Lowess regression at these levels of GDP per capita.
impose labor market and product market regulations to mitigate market failures, the failure of regulatory reform to keep pace with economic development can introduce barriers to structural transformation.
The other set of determinants involve market failures such as coordination problems that result from the high cost of collating and processing information on new products, technologies, and industries, and credit market failures. Both of these market failures involve externalities and information asymmetries.
Coordination problems typically arise when externalities create significantly large differences between private and social returns. In particular, a problem arises when externalities depress private returns. Investment in new technology, for example, is risky, with the first mover typically bearing the brunt of the risk. Meanwhile, successful investment invites the entry of other investors and undercuts the first mover’s effective rate of return on the investment. To avoid exposure to this risk, the investor may forego the investment unless it is possible to prevent or delay the entry of competition, even though society would view competition as beneficial. This state of affairs discourages investment in new technology and the adoption of new production processes unless governments intervene.
Credit market failures often stem from bad choices and moral hazards in the allocation of loans. Developing economies typically have underdeveloped financial markets and suffer severe information asymmetries. In addition, misguided government policies can force the misallocation of credit, making it difficult to obtain financing for new projects with promise and to scale up operations that have already proven their worth (León-Ledesma and Christopoulus 2016). The resulting low rate of investment and small scale of operations in high-productivity sectors slows the pace of structural transformation. To overcome credit market failures, governments can implement selective credit policies by providing credit directly to targeted sectors and indirectly through loan guarantees.
Macroeconomic determinants
of potential growth
The panel of Asian countries under study is augmented by including other economies, both emerging and advanced, to arrive at a total of 69 economies studied over the period 1960–2014 (the list is reported in Table A1 in the Appendix). As potential growth is the sum of the growth rates of potential labor productivity and the labor force, the analysis needs to account for both components. Moreover, the decomposition analysis indicates that potential labor productivity growth is attributed mostly to productivity growth within sectors.
The search for the determinants of productivity growth is complex because the literature provides a long list. To overcome this difficulty, various procedures for model selection were devised with the objective of determining which variables robustly correlate with economic growth (e.g., Sala-i-Martin et al. 2004). One such methodology is applied in this chapter: the Bayesian model-averaging approach to estimating
